introvoyz041/Guacamol

--- dataset_info: features: - name: SELFIE dtype: string - name: SMILES dtype: string splits: - name: train num_bytes: 351930666 num_examples: 1273077 - name: val num_bytes: 21949958 num_examples: 79564 - name: test num_bytes: 65952355 num_examples: 238694 download_size: 147769741 dataset_size: 439832979 configs: - config_name: default data_files: - split: train path: data/train-* - split: val path: data/val-* - split: test path: data/test-* --- # Dataset Card for Guacamol Dataset from the [Guacamol](https://github.com/BenevolentAI/guacamol) benchmark ([paper](https://arxiv.org/abs/1811.09621)). Dataset contains two columns, SMILE and SELFIE. Splits are identical to original splits, however, any SMILE that could not be converted to a SELFIE was dropped. Likewise, any SELFIE in the val/test splits that contained a token not found in the train split was dropped. Can be used with [this tokenizer](https://huggingface.co/haydn-jones/GuacamolSELFIETokenizer).